Potential conflict of interest: Nothing to report.
In our previous work, including analysis of more than 10,000 sera from control patients and patients with a variety of liver diseases, we have demonstrated that with the use of recombinant autoantigens, antimitochondrial autoantibodies (AMAs) are only found in primary biliary cirrhosis (PBC) and that a positive AMA is virtually pathognomonic of either PBC or future development of PBC. Although the mechanisms leading to the generation of AMA are enigmatic, we have postulated that xenobiotic-induced and/or oxidative modification of mitochondrial autoantigens is a critical step leading to loss of tolerance. This thesis suggests that a severe liver oxidant injury would lead to AMA production. We analyzed 217 serum samples from 69 patients with acute liver failure (ALF) collected up to 24 months post-ALF, compared with controls, for titer and reactivity with the E2 subunits of pyruvate dehydrogenase, branched chain 2-oxo-acid dehydrogenase, and 2-oxo-glutarate dehydrogenase. AMAs were detected in 28/69 (40.6%) ALF patients with reactivity found against all of the major mitochondrial autoantigens. In addition, and as further controls, sera were analyzed for autoantibodies to gp210, Sp100, centromere, chromatin, soluble liver antigen, tissue transglutaminase, and deaminated gliadin peptides; the most frequently detected nonmitochondrial autoantibody was against tissue transglutaminase (57.1% of ALF patients). Conclusion: The strikingly high frequency of AMAs in ALF supports the thesis that oxidative stress-induced liver damage may lead to AMA induction. The rapid disappearance of AMAs in these patients provides further support for the contention that PBC pathogenesis requires additional factors, including genetic susceptibility. (HEPATOLOGY 2007.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Environmental factors, in particular xenobiotics, have been implicated in the etiology of primary biliary cirrhosis (PBC).1–3 Specifically, it has been demonstrated that in addition to the native mitochondrial autoantigens, antimitochondrial autoantibodies (AMAs) also recognize xenobiotic modified E2 subunits of pyruvate dehydrogenase (PDC-E2).4 Several such candidate xenobiotics include chemicals found in detergents, food additives, and cosmetics.4–6 Interestingly, although immunization of experimental animals with xenobiotic modified E2 subunits induces true self-reacting AMA, it does not appear to induce clinical liver disease.7 Fulminant hepatic failure, also known as acute liver failure (ALF), is a clinical condition that reflects both oxidant stresses and hepatic damage. ALF is characterized by severe and sudden liver cell dysfunction leading to coagulopathy and hepatic encephalopathy in previously healthy persons with no known underlying liver disease.8 Although more than 43% of patients with ALF survive without a transplant, 28% die and 29% undergo liver transplantation.9
Hepatic insult secondary to acetaminophen (APAP) (paracetamol, N-acetyl-p-aminophenol) consumption has a major oxidant component; toxic doses of APAP produce reactive oxygen and nitrogen species and reactive metabolites in experimental animals or cultured cells.10, 11 Importantly, AMA induction and perpetuation appears sensitive to the oxidative state of the tissues involved.12–14 We hypothesized that in patients with ALF, AMA may be induced by the combination of oxidative stress and liver cell death, resulting in exposure of the immune system to both native and modified intracellular proteins. This transient induction of AMA could proceed to PBC if the immune response occurs in genetically susceptible subjects.15, 16
The objective of this study was to make use of serum samples of patients with ALF obtained by the Acute Liver Failure Study Group to investigate if and to what extent these patients' sera contain AMAs.17 In addition, we have also analyzed sera for autoantibodies against several additional autoantigens, including gp210, sp100, centromere, chromatin, soluble liver antigen (SLA), tissue transglutaminase (tTG), and deaminated gliadin peptide (DGP). We report the results of our analysis of autoantibody reactivity of 217 serum samples in 69 subjects with ALF collected post-ALF (days 1-7) and at 12 months and 24 months follow-up.
The Acute Liver Failure Study Group comprises 23 sites around the United States that have developed a prospective registry of cases of ALF of all causes since 1998 (for more information and a list of the sites, see www.acuteliverfailure.org). More than 1,100 patients have been enrolled with detailed clinical information; serum and DNA data are available for most patients. All patients enrolled meet criteria for encephalopathy (any grade) and coagulopathy (international normalized ratio ≥1.5). In this study, 217 serum samples from 69 subjects (46 females, 23 males) with a clinical diagnosis of ALF were selected to represent the standard causes associated with ALF; for the most part, samples were obtained during initial hospitalization, at 12 months post-ALF, and at 24 months post-ALF. Of the 217 serum samples, 176 samples were obtained between days 1 and 7 (86 samples between days 1 and 4; 90 samples between days 5 and 7) for all 69 subjects, 27 samples were obtained at 12 months follow-up, and 14 samples were obtained at 24 months follow-up. The causes of ALF included APAP (n = 26), drug-induced liver injury (n = 11), autoimmune hepatitis (n = 9), hepatitis A virus infection (n = 3), hepatitis B virus infection (n = 4), indeterminate (n = 6), and miscellaneous (n = 10). Sera from patients with PBC (n = 15) and healthy volunteers (n = 81) from our laboratory sera banks were used throughout as positive and negative controls as detailed previously.17
Recombinant glutathione fusion proteins containing the AMA epitopes of PDC-E2, BCOADC-E2, OGDC-E2, and MIT3 (a triple hybrid containing the immunodominant epitopes of PDC-E2, BCOADC-E2, and OGDC-E2) were cloned and expressed in pGEX-4T-1. Briefly, a 414-bp EcoRI fragment coding for the PDC-E2 amino acid residues 91-228 which encodes the inner lipoyl domain and part of the outer lipoyl domain was amplified by polymerase chain reaction, purified and cloned into the EcoRI site of pGEX-4T-1, transformed and expressed in Escherichia coli DH5α. Likewise, the BCOADC-E2 epitope (amino acid residues 1-118) and OGDC-E2 epitope (amino acid residues 67-147) were amplified via polymerase chain reaction, cloned, and expressed in the BamHI site and NotI site, respectively, of pGEX-4T-1.17, 18 Successful cloning and expression were confirmed via DNA hybridization and immunoassay using monoclonal antibodies to PDC-E2, BCOADC-E2, and OGDC-E2.19 An irrelevant control Met e I in plasmid pGEX-4T-1 was also purified as described previously.20
Detection of AMA.
All assays for AMA were performed blindly at the University of California, Davis (by P. L.), and Teikyo University (by H. M.) using similar methodology. Sera from subjects with ALF and controls were diluted 1:1,000 and analyzed for AMA reactivity against the triple hybrid MIT3.17 ALF sera that were positive to MIT3 were absorbed against an irrelevant protein control, Met e I,20 then retested for reactivity against MIT3 and Met e I via immunoblotting. Briefly, recombinant MIT3 and Met e I were resolved individually via SDS-PAGE at 30 mA constant current. SDS-PAGE–resolved proteins were transferred to nitrocellulose filters (Micron Separations Inc., Westboro, MA) and cut into 3-mm strips, blocked with 3% milk in phosphate-buffered saline (PBS) and probed with sera from patients with ALF for 1 hour at room temperature. After 3 10-minute washes with phosphate-buffered saline/0.05% Tween 20 (phosphate-buffered saline/Tween), the filters were incubated with horseradish peroxidase–conjugated anti-human Ig, washed with phosphate-buffered saline/Tween, and developed via chemiluminescence using the Supersignal West Pico Chemiluminescence Substrate (Pierce, Rockford, IL). Sera from AMA-positive PBC patients and AMA-negative controls were used as controls throughout. Sera that were positive to MIT3 after absorption but not the irrelevant protein Met e I were scored as AMA-positive. Additionally, all ALF sera were assayed for reactivity against recombinant proteins of PDC-E2, BCOADC-E2, and OGDC-E2 at 1:500 dilution via immunoblotting as described previously.17
AMA-positive samples were assayed for titer to recombinant mitochondrial autoantigens via ELISA as described previously.21 Positive AMA reactivity is defined as more than 3× OD 405 ± SD than that of healthy volunteers.
Detection of Autoantibodies Against gp210, Sp100, Centromere, Chromatin, SLA, tTG, and DGP.
IgG to gp210, Sp100, centromere, chromatin, SLA, tTG, and DGP were determined using ELISA kits (INOVA Diagnostics, San Diego, CA) to each of the autoantigens. Briefly, precoated ELISA plates were incubated with 100 μl of diluted patient sera (1:100) for 30 minutes at room temperature and then washed 3 times with horseradish peroxidase wash buffer. The plates were then incubated with 100 μl of horseradish peroxidase–conjugated goat anti-human IgG for 30 minutes and washed 3 times with horseradish peroxidase wash buffer. Reactivity was detected by incubating with 3,3′,5,5′ tetramethyl benzidine (TMB) in the dark for 30 minutes. The reaction was terminated by the addition of 100 μl of stop solution and the absorbance read at 405 nm. In all assays, low positive, high positive, and negative controls were included.
Autoantibody reactivity between ALF sera and normal sera against MIT3, gp210, Sp100, centromere, chromatin, SLA, tTG, and DGP were analyzed via Student t test. The frequency of AMA-positive sera between ALF sera and the estimated prevalence of PBC were compared statistically via Fisher exact test with SAS version 9 software. The statistical differences between male and female ALF subjects were analyzed via Student t test. A P value of less than 0.05 was considered statistically significant.
Serological Autoantibody Profile of ALF and Healthy Volunteers.
In ALF subjects, the most frequently detected autoantibodies were directed against tTG and MIT3, with 36/69 (57.1%) and 23/69 (33.3%) of ALF subjects reacting to tTG and MIT3, respectively. Only 1/69, 6/69, 3/69, 5/69, 1/69, 36/69, and 8/69 of the ALF subjects reacted with gp210, Sp100, centromere, chromatin, SLA, tTG, and DGP, respectively (Table 1). Only 0/81, 0/81, 2/81, 3/81, 0/81, 1/81, 1/81, and 3/81 of the normal sera reacted with MIT3, gp210, Sp100, centromere, chromatin, SLA, tTG, and DGP, respectively.
Table 1. Autoantibody Reactivity of ALF Sera
No. of Reactive ALF Subjects (%)
No. of Reactive Normal Subjects (%)
P < 0.0001 between ALF and control subjects, determined via Student t test. †Reactivity to tTG was only transient. Only 1 subject remained tTG-positive at 24 months after ALF.
Detection of AMA via Immunoblotting.
AMA-positive serum samples from subjects with ALF have the same AMA specificity as observed in sera from patients with PBC to recombinant PDC-E2, BCOADC-E2, and OGDC-E2 via immunoblot (Fig. 1). AMAs were detected in 28/69 (40.6%) ALF subjects. Serum samples from 22/69 and 20/69 subjects collected from days 1-4 and days 5-7, respectively, were found to have AMA against 1 or more mitochondrial autoantigens. Eight of 27 subjects had AMA at 12 months post-ALF, including 2 subjects who were not AMA-positive until 12 months follow-up. One subject remained AMA-positive at 24 months follow-up (Table 2). At 1-4 days post-ALF, 18/69, 6/69, and 7/69 subjects reacted against PDC-E2, BCOADC-E2, and BCOADC-E2, respectively. At 5-7 days post-ALF, 16/69, 5/69, and 6/69 subjects reacted against PDC-E2, BCOADC-E2, and BCOADC-E2, respectively. At 12 months post-ALF, 4/27, 2/27, and 4/27 of ALF subjects reacted against PDC-E2, BCOADC-E2, and BCOADC-E2, respectively. Only 1/14 subjects remained AMA-positive at 24 months post-ALF and reacted against OGDC-E2.
Table 2. Time Course of AMA Reactivity and Antigen Specificity to PDC-E2, BCOADC-E2, and OGDC-E2
Time of Sample Collection
No. of AMA-Positive Subjects/ Total No. of Subjects*
No. of Subjects Reacted With PDC-E2
No. of Subjects Reacted With BCOADC-E2
No. of Subjects Reacted With OGDC-E2
Two subjects who were previously AMA-negative became AMA-positive at 12 months.
Antigen Specificity and Time Course of AMA Reactivity in AMA-Positive ALF Sera.
AMA-positive ALF subjects reacted to the major mitochondrial autoantigens including PDC-E2, BCOADC-E2, and OGDC-E2 in PBC (Tables 2 and 3). Among the 86 serum samples obtained from days 1-4 post-ALF, 28/86 reacted with PDC-E2, 10/86 reacted with BCOADC-E2, and 12/86 reacted with OGDC-E2. Among the 90 serum samples obtained from 5-7 days post-ALF, 25/90 recognized PDC-E2, 9/90 recognized BCOADC-E2, and 10/90 recognized OGDC-E2. Of the sera samples obtained from 27 subjects obtained at 12 months follow-up, 4/27, 2/27, and 4/27 remained reactive to PDC-E2, BCOADC-E2, and OGDC-E2, respectively. Only 1/14 serum samples obtained at 24 months post-ALF was AMA–reactive, and it recognized OGDC-E2 only (Table 2). PDC-E2 is the predominant antigen, while a similar frequency of AMA reactivity against BCOADC-E2 and OGDC-E2 were detected from the sera samples. Sixteen subjects recognized PDC-E2 only, 3 recognized both PDC-E2 and BCOADC-E2, while 7 subjects recognized PDC-E2, BCOADC-E2, and OGDC-E2. Six AMA-positive ALF subjects were found to react with OGDC-E2 only. One subject was positive to both BCOADC-E2 and OGDC-E2. None reacted with BCOADC-E2 only (Table 3).
Table 3. Antigen Specificity of ALF Sera to Mitochondrial Autoantigens
Determined by immunoblotting against recombinant mitochondrial autoantigens.
PDC-E2 and BCOADC-E2
PDC and OGDC-E2
BCOADC-E2 and OGDC-E2
PDC-E2, BCOADC-E2, and OGDC-E2
The titers to PDC-E2 ranged from 1:80 to 1:2,560 in samples obtained from days 1-4, and 1:80 to 1:1280 in samples obtained from days 5-7 and at 1:160 to 1:640 at 12 months post-ALF. The titers to BCOADC-E2 ranged from 1:640 to 1:2,560 in samples obtained from days 1-4 and at 1:160 to 1:1280 in samples obtained from days 5-7. Two serum samples remained reactive to BCOADC-E2 at 12 months post-ALF with titers of 1:640 and 1:1,280. The titers to OGDC-E2 ranged from 1:80 to 1:2,560 in samples obtained from days 1-4 and at 1:80 to 1:1,280 at days 5-7. Four serum samples remained reactive to OGDC-E2 at 12 months and 24 months with a titer of 1:80 to 1:1,280. One serum sample was reactive to OGDC-E2 with a titer of 1:640 even at 12 months post-ALF (Fig. 2).
AMA Is Not Limited to Any Specific ALF Diagnosis and Is Detected in Both Male and Female ALF Subjects.
Though limited by the small sample size in each group of ALF subjects, our data do not relate the detection of AMA to any particular causal agent, because AMAs were detected in 9/26 of APAP, 5/11 of drug-induced liver injury, 4/9 of autoimmune hepatitis, 2/3 of hepatitis A patients, 1/4 of hepatitis B patients, and 4/10 subjects with ALF due to other liver diseases. In addition, 3/6 subjects with unknown causes of liver failure were also found to be AMA-positive (Table 4). AMA reactivity was not different between female (20/46 [43.3%]) and male (8/23 [34.7%]) ALF subjects. The absence of a sex difference in AMA is striking as the female/male ratio in PBC approaches 9:1.
Table 4. Diagnosis and Sex Analyses of Acute Liver Failure Subjects With Positive AMA Reactivity
Total No. of AMA-Positive Subjects/Number of Subjects (%)
Expressed as number of reactive samples/total number of samples. †Includes known causes such as pregnancy-related, cancer, heat stroke, shock, non–hepatitis A or non–hepatitis B viruses. ‡Clinical cause unknown.
Drug-induced liver injury
Prevalence of AMA in ALF Subjects Compared With Prevalence of PBC.
The frequency of AMA-positive sera in our sample of ALF subjects is much higher than what is expected from the normal population. For example, the 95% CI for the frequency of AMA positivity in the group of ALF patients studied is 0.2895-0.5213 (P < 0.0001); this is significantly higher than the prevalence of PBC AMA, which varies from 19.1 per million to 402 per million.22, 23 Similarly, the 95% confidence limits for the frequency of AMA-positive subjects/total number of subjects at 1-7 days, 12 months, and 24 months are 0.2621-0.4908, 0.1237-0.4681, and -0.0639-0.2059, respectively (P < 0.0001).
Our laboratory has had extensive experience with the use of recombinant autoantigens. In studies of AMA reactivity in large numbers of sera, we have demonstrated that when recombinant antigens are used for immunodiagnosis, they are strikingly specific for PBC; we have also shown that such specificity and sensitivity is dramatically increased with the use of ELISA and/or immunoblot assays compared with immunofluorescence.21, 24, 25 In our initial analysis and our observations that AMAs were indeed found in ALF, we decided to replicate the data in a second site—hence the duplicate analysis at both UC Davis and Teikyo University (both sites had identical results). Recent work has suggested that posttranslational modification of the mitochondrial 2-OADC lipoyl moiety is a possible mechanism in the initiation of AMAs in PBC.6, 26 The 2-OADC lipoyl domains are structurally exposed on the periphery of the molecule and mobile, resembling a swinging arm that facilitates electron transfer. In light of the structural configuration of the lipoyl domains in the 2-oxo-acid dehydrogenase complex and the biochemical characteristics of lipoic acid (i.e., its ability to alternate between oxidized and reduced state in its electron transport function), 2-OADC E2s are highly susceptible to chemical modification through endogenous or exogenous influence that may affect the redox state of the cells.13
In this study, we analyzed 217 individual serum samples obtained from 69 patients with ALF in which catastrophic damage to hepatocytes and presumably bile duct epithelial cells has occurred. In this group of ALF patients, 28/69 subjects developed AMA to 1 or more of the mitochondrial autoantigens in PBC; only 1 subject remained AMA-positive at 24 months post-ALF. It is not known whether this subject developed PBC, because no further samples were available for testing. Although an extensive long-term follow-up program is in place, to date we have had difficulty finding patients who do not have a long-term illness except those who have received transplants. One of the limitations, of course, is that there is no follow-up on patients who die. In addition, most APAP cases cannot be located, because they are seldom found at the same address after 1 year.
AMA reactivity to each of the mitochondrial autoantigens remained rather constant from day 1-7, but decreased sharply by 12 months post-ALF (Table 3) and only 1 subject remained positive to a single mitochondrial autoantigen (OGDC-E2) by 24 months post-ALF. The rapid induction of AMA in ALF subjects suggests that liver injury can trigger the transient production of AMA. The AMA titer in AMA-positive ALF subjects ranged from 1:80 to 1:2,560 to PDC-E2, 1:160 to 1:2,560 to BCOADC-E2, and 1:80 to 1:2,560 to OGDC-E2 via ELISA, which is significantly lower than the AMA titer usually observed in patients with PBC.27 Clearly, the frequency of AMA-positive sera among ALF subjects in this study is many logs higher than those observed in the general population.
Of interest to the data herein is our inclusion of autoimmune hepatitis. Such cases were included primarily as an additional group for study. However, the percentage of AMA-positive subjects from the autoimmune hepatitis group is similar to the other included ALF patients. Autoantibodies from autoimmune hepatitis subjects do not appear to be directly involved in liver injury prior to ALF in AMA-positive autoimmune hepatitis subjects. Although there is no clear evidence of a higher incidence of AMA among various ALF diagnoses, it is interesting to note that AMA was found in almost 35% of APAP poisoning subjects. APAP overdose is the most frequent cause of ALF.28 APAP toxicity involves a major oxidant component generating reactive oxygen species.10, 11 Furthermore, the generation of reactive oxygen species is associated with mitochondrial damage in murine hepatocytes incubated with APAP.29 Our data indicate that AMAs are found in sera in almost 35% of APAP poisoning subjects, suggesting that the 2-OADC E2 lipoyl domain is a likely target of APAP-induced reactive oxygen and nitrogen species. It is known that a decrease in glutathione and CYP2E1 are risk factors in toxic liver injury, and both CYP2E1 and glutathione play important roles in APAP metabolism.30, 31 Human HepG2 cells that overexpress CYP2E1 have been shown to have elevated glutathione levels for protection against oxidative stress,32 and a recent study in patients with alcohol-associated hepatitis has shown that both glutathione and CYP2E1 messenger RNA were reduced by 70%-80%.33 The effect of APAP on the level of glutathione is of importance in the context of this study, because the lipoyl domain of PDC-E2 is modified by glutathionylation during apoptosis with this modification blocking PDC-E2 recognition by AMAs.12 This phenomenon is significant in PBC, because biliary epithelial cells do not glutathionylate PDC-E2 during apoptosis, whereas other liver cells do.12 Although it appears to be a logical assumption that APAP severity may be correlated with glutathione depletion and hence a higher titer of AMA by presentation of available nonglutathionylated PDC-E2 to the immune system, the titers of AMA observed did not correlate with disease severity in subjects with APAP. However, it is possible that there is a threshold of antigen involved in the breaking of tolerance and subsequent disease development. Similarly, although our data point to a role for oxidative stress in inducing AMAs, we do not wish to imply that oxidative stress is an absolute necessity for the pathophysiology of PBC. The patients studied reflect an extreme example of hepatic inflammation and oxidative stress; therefore, multiple pathways may be involved in the generation of autoantibodies. Similarly, it should be noted that in addition to APAP-induced liver injury, oxidative stress-induced injury is implicated in other ALF diseases.34–37 In this study, in addition to AMA, several other autoantibodies are detected in ALF patients. Thus, it remains a possibility that the induction of autoantibodies is an epiphenomenon in patients with ALF.
In conclusion, the demonstration of the high frequency of AMA induction by ALF, albeit transient, provides putative associations of oxidative damage in the pathophysiology of PBC. This result, along with the absence of any sex-related differences demonstrated by this study, suggests that the perpetuation of the AMA response and development of PBC will be a multifactorial process that would include an oxidative insult in the context of a specific accompanying etiological scenario comprised of genetic susceptibility and appropriate innate and adaptive immune responses.38 Future studies should address this issue by studying the administration of drugs that induce ALF in the recently described murine models of human PBC.39–41
Members and institutions participating in the Acute Liver Failure Study Group: W. M. Lee, M.D., Julie Polson, M.D., Carla Pezzia, University of Texas Southwestern, Dallas, TX; Anne Larson, M.D., University of Washington, Seattle, WA; Timothy Davern, M.D., University of California, San Francisco, CA; Paul Martin, M.D., Mount Sinai School of Medicine, New York, NY; Timothy McCashland, M.D., University of Nebraska, Omaha, NE; J. Eileen Hay, M.D., Mayo Clinic, Rochester, MN; Natalie Murray, M.D., Baylor University Medical Center, Fort Worth, TX; A. Obaid Shakil, M.D., University of Pittsburgh, Pittsburgh, PA; Andres Blei, M.D., Northwestern University, Chicago, IL; Atif Zaman, M.D., University of Oregon, Portland, OR; Steven Han, M.D., University of California, Los Angeles, CA; Robert Fontana, M.D., University of Michigan, Ann Arbor, MI; Brendan McGuire, M.D., University of Alabama, Birmingham, AL; Ray Chung, M.D., Massachusetts General Hospital, Boston, MA; Alastair Smith, M.B., Ch.B., Duke University Medical Center, Durham, NC; Michael Schilsky, M.D., Cornell/Columbia University, New York, NY; Adrian Reuben, MBBS, Medical University of South Carolina, Charleston, SC; Santiago Munoz, M.D., Albert Einstein Medical Center, Philadelphia, PA; Rajender Reddy, M.D., University of Pennsylvania, Philadelphia, PA; R. Todd Stravitz, M.D., Virginia Commonwealth University, Richmond, VA; Lorenzo Rossaro, M.D., University of California, Davis, Sacramento, CA; Raj Satyanarayana, M.D., Mayo Clinic, Jacksonville, FL; and Tarek Hassanein, M.D., University of California, San Diego, CA.